1122
FRANK K.
SIGNAIGO AND
HOMER ADICINS
Vol. 55
[A CONTRIBUTION FROM THE LABORATORY O F ORGANIC CHEMISTRY OF THE UNIVERSITY O F WISCONSIN]
A Synthesis of dl-Proline from Pyrrole' BY FRANK K. SIGNAIGO AND HOMER ADKINS Several syntheses of dl-proline have been reported but none of them is very satisfactory from a preparational standpoint. The best methods are apparently those of Heymons,2 who obtained the copper salt of dl-proline in three operations from piperidone-2, and that of Sgrensen and Anderseq3 who prepared the copper salt in five operations from malonic ester. The first method is said to give an over-all yield of 28%. The yield by the second method is not given, but is apparently somewhat greater. Several authors have suggested the desirability of developing a synthesis of proline from pyrrole through 2carbethoxypyrrole as by P ~ t o k h i n . ~A difficulty in this process lies in the hydrogenation of the pyrrole ring in a carboxy- or carbethoxypyrrole. Putokhin used a high ratio of Adams platinum catalyst with 1 g. of pyrrole-2-carboxylic acid and as a result of repeated treatments with hydrogen and catalyst and change of solvent, obtained a 23% yield of the copper salt of dl-proline from the pyrrole acid. While the hydrogenation of a 2-carboxy- or carbethoxypyrrole seems an impractical process with the catalysts now available, it appeared feasible to prepare dl-proline through the formation of 1,2-dicarbethoxypyrrole and the hydrogenation of the latter, i. e . CH-CH
I/
CH
11
CH
\N& CHz-CHz
"
PYRRYLMAGNESIUY BROMIDE
CH-CH EtMgBr
A
ClCOOEt
11
CH
11
0'
CCOOEt
\/
N-COOEt NiJHz CHz-CHz
I 1 HzO 1 I CHz CHCOOH tCH2 CHCOOEt \/
However, by this process crystalline dl-proline has been obtained in an over-all yield of 57% from pyrrole. Odd06 prepared 2-carbethoxypyrrole in over 90% yield from pyrrylmagnesium bromide and ethyl chlorocarbonate. In attempting t o follow his method, but on a larger scale, we recovered 42% of the pyrrole and obtained 7% of l-carbethoxypyrrole, 22% of 2-carbethoxypyrrole and 2% of 1,2-dicarbethoxypyrrole. When the conditions were modified by heating the reaction mixture for four hours at 35' before hydrolysis, instead of keeping it near 0') then the yield of 1,2-dicarbethoxypyrrole was much greater, i. e., 26%) accompanied by 13% of 1-carbethoxypyrrole and 3% of 2-carbethoxypyrrole. However, 50% of the pyrrole was recovered. These results suggested that pyrrylmagnesium bromide reacts with ethyl chlorocarbonate to give both 1- and 2-carbethoxypyrrole and that the latter but not the former reacts with pyrrylmagnesium bromide to give pyrrole and 2-carbethoxypyrrylmagnesium bromide. The last mentioned compound may then react, especially a t higher temperatures, with ethyl chlorocarbonate to give 1,2-dicarbethoxypyrrole. These transformations are indicated below.
\/
N-COOEt
The hydrogenation may be carried out easily and almost quantitatively as a result of work recently reported.6 The hydrolysis step is also a simple process but the preparation of the 1,2-dicarbethoxypyrrole in a good yield required study. (1) The Wisconsin Alumni Research Foundation allotted funds for a research assistantship held by F. K.S. in the period Sept., 1933 to Sept., 1935. (2) Heymons, B e y . , 66, 846 (1933). (3) Spensen and Andersen, Z . physiol. Chem., 66, 240 (1908). (4) N. Putokhin, J . Russ. Phys.-Chcm. Soc., 62, 2216 (1930); C . A , , 96, 3995 (1931). (6) Signaigo and Adkins, THIS JOURNAL, 68,709 (1936).
1
ClCOOC2Hs
I I
HC-CH HC /I CH ll
4
HC-CH 'Hk
kCOOC2Ha
+ MgBrCl
H
COOCIH~
Pyrr ylmagnesiurn bromide O0/ HC-CH HzO HC-CH Pyrrole /I II 11 I/ HC CCOOC~H, H C CCOOCzHr,
Y H
'0s
+
'4
MgBr C1COOCzH6~35' HC-CH HC I1 CCOOCzHr I/ \/
(6) Oddo, Gnee. chim. i t d , 39, I, 649 (1909); 42, 11,282 (1812).
A SYNTHESIS OF &PROLINE FROM PYRROLE
July, 103G
Upon the basis of this scheme i t seemed desirable to treat one mole of ethylmagnesium bromide with one mole of pyrrole and then to add nlternafely in portions, one mole each of ethyl chlorocarbonate and ethylmagnesium bromide and finally a second mole of ethyl chlorocarbonate. By such a process the reaction of all of the pyrrole was brought about and a t the same time excessive reaction between the ethyl chlorocarbonate and the ethylmagnesium bromide was avoided. Such a procedure as described in the experimental part gave a 65% yield of l12-dicarbethoxypyrrole. There seem to be two rather reasonable explanations of the fact that 1- and 2-carbethoxypyrroles are formed simultaneously. Tschelinzew and Karmanov’ suggested that the isomerism of the pyrrole nucleus makes possible the formation of two “pyrrylrnagnesium bromides,” i. e. HC-CH
+ CJ-LMgBr +HC-CHJH + C2Hs
v NH
\/
N-MgBr HC=CH
HC=CH
The formation of a Grignard reagent in the second case would be analogous to the behavior of a methylene group in cyclopentadiene, indene and fluorene. The formulation is in accord with the fact that only pyrroles which may tautomerize form Grignard reagents? If the above is accepted then the relative proportions of 1- and 2-carbethoxypyrroles formed are a function of the proportion of the tautomers and (or) of the relative rates of reaction of ethyl chlorocarbonate with the two Grignard reagents. HC-CH
1
Hk
(1
CH
+ ClCOOCzHs
\/
N-MgBri
HC-CH (H&
&COOCzHs + Hcl)
v N-MgBr
HC-CH
I1 I/
HC
I
J.
CCOOEt
+ MgClBr
\&I (7) (a) Tschelinzew and Karmanov, J. Russ. Phw-Chem. Soc., 47, 161 (1915); (b) cf. Oddo, Gazz. chim. i l d . , 64, 584 (1934). (8) O d d o , ibid., 44, I, 706 (1914).
1123
However, it is not necessary to assume the existence of tautomeric pyrroles in order to explain the formation of 2-carbethoxypyrrole, for the ethyl chlorocarbonate may condense with the hydrogen on the 2-carbon atom. Such a scheme finds support in the fact that 1methylpyrrole does not form a Grignard reagent yet condenses with acid chlorides in the presence of an alkylmagnesium halide to yield l-methyl-2acylpyrroles.8 However I-carbethoxypyrrole apparently did not react with ethyl chlorocarbonate in the presence of a Grignard reagent. Experimental Part The apparatus for the preparation of 1,2-dicarbethoxypyrrole consisted of a 1-liter, 4-necked flask fitted with a mercury-sealed stirrer, a reflux condenser, the upper end of which was connected to a drying tube containing calcium chloride, and two dropping funnels. One of these dropping funnels was of the usual type, and the other was a 500-cc. flask with a stopcock sealed on the bottom and so fitted that a Grignard reagent could be prepared in it. Both of the funnels were graduated. Ethylmagnesium bromide (approximately 0.25 mole) was prepared in the reaction flask from 32.7 g. (0.30 mole) of ethyl bromide, 7.2 g. (0.30 mole) of magnesium turnings and 150 cc. of ether. A like amount also was prepared in the 500cc. dropping funnel. The reaction flask was cooled with a mixture of ice and salt t o below 0” and 17 g. (0.25 mole) of pyrrole in 35 cc. of ether was added dropwise through the smaller dropping funnel. The reaction mixture was stirred during this and subsequent operations. Into the smaller dropping funnel was placed 75 cc, of an ether solution containing 28 g. (0.25 mole) of ethyl chlorocarbonate. The Grignard reagent in the larger dropping funnel was diluted with ether to a volume of 200 CC. These two reagents were then added to the reaction flask drop by drop in the following fashion. One-half of the ethyl chlorocarbonate solution (37 cc.) was f i s t added. This was followed by one-half of the Grignard solution (100 cc.). Again half of the remaining ethyl chlorocarbonate solution (19 cc.) was added and followed by half of the remaining Grignard solution (50 cc.). This process was repeated, each time adding one-half of the solution remaining in the dropping funnel until only 5-10 cc. of each reagent remained. This last small amount of each solution was then added and 37 g. (0.35 mole) of ethyl chlorocarbonate was poured into the reaction mixture through the smaller dropping funnel. These addition processes were completed in two hours. The reaction mixture was allowed to warm up to room temperature and then refluxed on a steam-bath for four hours. The contents of the reaction flask were then poured onto ice and the mixture made slightly acid by the addition of dilute sulfuric acid. The ether layer was separated and the aqueous layer was extracted twice with ether. The combined ether solutions were then washed once with a small quantity of sodium bicarbonate solution and dried for eight hours over anhydrous sodium sulfate. The ether was distilled off and the liquid residue was fractionated a t reduced pressure as indi-
1124
W. K. MCEWEN
cated in the table. The recovery of ethyl chlorocarhonate as well as the amount of I-carbethoxypyrrole obtained, indicate that 1.75 mole each of the Grignard reagent and the ethyl chlorocarbonate per mole of pyrrole would give as good a yield as did the proportion of reactants used above.
TABLE I PRODUCTS OF
Fraction
Compound
REACTION Weight, Yield,
Boiling range, OC
6.
1 Ethyl chlorocarbonate Up to 55' (65 mm.) 2 I-Carbethoxypyrrole 60- 65" (7 mm.) 3 2-Carbethoxypyrrole 100-110' (7mm.) 4 l,2-Dicarbethoxypyrrole 125-135' (7 mm.) Residue
10
%
..
9 . 2 26.4
0.5
1.5
34.2 65.0 1.8 Total 92.9
The fourth fraction was found to be nearly pure 1,2dicarbethoxypyrrole6 since on refractionation 98% of the material distilled at 132-134' a t 8 mm. The l-carbethoxypyrrole had a b. p. of 57-58' a t 7 mm., and 178-180' a t 740 mm., and d26 1.037, n% 1.4898. The boiling point of this compound checks with that reported by other investigators. Only one author has reported the density and refractive index of this compound and these values do not check with those given above. However, the compound seems t o be sufficiently characterized by the fact that it takes UP 4 atoms of hydrogen to give a non-basic cornpound having the correct analysis for a carbethoxypyrrolidine, indicating that the carbethoxy group is on the nitrogen atom. The 1-carbethoxypyrrole is better prepared from potassium pyrrole and ethyl c h l o r ~ a r b o n a t e . ' ~ ' ~ The 2 - c a r b e t h o x ~ ~ ~ r r had o l e a m. P. of 41-42' when (9) Ciamician and Dennstedt, Ber., 16, 2579 (1882); Bet'., 60, 196 (1927).
[CONTRIBUTION FROM
THE
Vol. 58
crystallized from low-boiling petroleum ether and a b. p. of 102-104' at 7 mm. The 1,2-dicarbethoxypyrrole so obtained was hydrogenated rapidly and quantitatively over Raney nickel.' In a typical experiment 1,2-dicarbethoxypyrrolidineb. p. 133-134' (8 mm.) was obtained in 98% yield by the hydrogenation of 0.2 mole of the pyrrole ester over 5 g. of Raney nickel under 100 to 200 atm. of hydrogen a t 50 to 85' in less than one hour. The 1,Z-dicarbethoxypyrrolidine(10.6 g. or 0.05 mole), was refluxed with 25 cc. of concentrated hydrochloric acid for three hours; 10 cc. more of hydrochloric acid was then added and the refluxing continued for two hours. The mixture was then evaporated on a steam-bath. The partially crystallized residue was dissolved in 100 cc. of water and treated with 0.10 mole of freshly precipitated silver carbonate. After filtration, the filtrate was treated with hydrogen sulfide. The solution was heated with I g. of activated charcoal and this together with the silver sulfide was removed by filtration. Upon evaporation of the filtrate on a steam-bath, there was obtained 5.7 g. of a gummy crystalline mass. This was dissolved in 25 cc. of hot ethanol and 50 cc. of hot dioxane was added. After cooling the solution with ice, the crystalline proline was filtered off arid dried in a desiccator. The yield was 5.2 g. (go%), m. p. 204'. The picrate was prepared by dissolving 1g. of proline and 2 g. of picric acid in 2 cc. of hot acetic acid, followed by addition of 7 cc. of ether and cooling. The picrate was recrystallized from ethanol and showed a m . p. of 136'.
Summary Crystalline &proline may be prepared in yields of 55 to 60% from pyrrole through the intermediate formation Of 1~2-dicarbethoxyp~ro1e and lj2dicarbethoxypyrrolidine. A method for the synhas been devised thesis of 1,2-dicarbethoxypyrrole . -. and developed. MADISON, WISCONSIN
RECEIVED MARCH13, 1930
CONVERSE MEMORIAL LABORATORY O F HARVARD UNIVERSITY]
A Further Study of Extremely Weak Acids BY W. K. MCEWEN The purpose of this paper is to report the new methods used and the results obtained in the study of extremely weak acids. It has been shown by Conant and Wheland' that reactions of the type RIHf R2RI- + RzH may be used to determine the relative acid strength of very weak acids such as the aromatic and aliphatic derivatives of water, ammonia and methane. The determination of the relative (1) Conant and Wheldnd, THISJ O U R N A L , 64, 1212 (1032).
amounts of the different substances present a t the point of equilibrium is a measure of the relative strength of the two acids. This can be determined in either of two ways. The first method used was to observe the change in color. Since many compounds of the type Rare colored, it is only necessary to mix the reactants and observe whether or not a change in color results. As a check the products of the reaction are also mixed. From such observations the compounds can be arranged in a series of de-
